Precision Recall scores doesn't match visualization-CodePudding

I am calculating the Precision Recall score for 3 different models. For my first model, the plot does not match the auc score which is 0.85 but it should be 0.5.

enter image description here

Here is my code:

precision, recall, thresholds = precision_recall_curve(y_test, pred1)
print(pred1)
fig, ax = plt.subplots()

ax.plot(recall, precision, color='red')

ax.set_title('Precision-Recall curve')
ax.set_ylabel('Precision')
ax.set_xlabel('Recall')
plt.show()

auc_precision_recall = auc(recall, precision)

CodePudding user response：

Sorry I don't have enough rep to comment.

Notice your y-axis, it is between 0.7-1, so an AUC of 0.85 seems to match the plot. The AUC would be 0.5 if the y-axis ranges from 0-1.

You can set the y-axis to be between 0-1 with ax.set_ylim((0,1))